Andi - what was the battery configuraton and batteries used? On 4-5 SinkPAD builds of XM-L2's and U3 1C I've seen low amps on standard 1 battery unprotected Panasonics. It's been very consistent on Nanjg or KD V2 drivers. The 26650 ICR KK does better, and 3 cells in parallel seemed to do better.
thats good to know - i am going to be putting an XML2 in my defiant 3c with 2x 18650 ncrb protected cells…. im going to be using a 3.5a constant current buck driver as well
I find this example useful as it does not compare MCPCBs at different currents. This one seems fair to say that gains are negligible with this SinkPad tech.
So it's XM-L(1)T6 vs XM-L(1) U3 with the same driver same cells and charged at the same levels (I know "fresh" does not extractly mean identical voltage with the other test)
According to specs we can average XM-L T6 @ 3A to output 943 lumens and XM-L U3 @ 3A to output 1072 lumens. So U3 should output on average 13.5% more light.
Now let's compare the 16mm aluminum MCPCB (thickness not specified, there are both 1mm and 1.6mm available on the market) with the 20mm (cut to unspecified diameter) copper MCPCB from SinkPad.
So to take those gains for the U3 (Cu) compared to T6 (Al) we have at start 9% output increase and at 30 seconds a 11.5% output increase. So that is under the average that the U3 should output compared to T6. Therefore I find that the copper had not much influence on the results. Simply by changing the T6 to U3 a bigger difference could theoretically be achieved (no copper involved).
Now it's a bit too late to make the output measurements at 3 minutes of runtime on Al vs Cu at the same ambient temperature.
Right - initial setup was a U3 1C on alum star, original alum pill.
Unfortunately I've just been grabbing whatever is fresh off the charger, but sometimes I see no difference in 2900 vs 3400, sometimes the 2900 seem better, or the 3400 seems better. Also, yes, a lot of things were changed: emitter, star, pill, driver. I think the HD 2010 before/after is one of the most impressive improvements and only changed the U3 1C/16mm alum star to a XM-L2 U2/20 mm SinkPAD. I think the ICR 26650 KK helped there as well by getting more out of the combo.
as Relic stated, differences in output = emitter swap, differences in heat sag = star. Alu star = 4.7% drop after 30s, Cu star = 2.1% drop after 30s. That might be just the consequence of the extra thermal mass of the Cu star (they’re relatively very heavy), it’s hard to know at just 30s.
If you really wanted to know the effect of the sinkpad stars on output after a set amount of time, you’d really have to compare the same LED on both types of star, by reflowing it from one to the other (and back again, to make sure). I doubt many people really have the inclination to go to that bother and even if they did, the sample sizes are likely to be small and the measurement/ experimental error large.
I think you can safely say that the individual contribution of a sinkpad to a light’s output is going to be small and reliant on an awful lot of different factors, but in combination with a well lapped star>pill interface, a good pill>housing interface and a good LED it will certainly help. We’re not talking big differences here, but every little bit helps!
An fyi.. Panasonic PD's, and better still Samsung INR 1500's do much better than Panasonic 3400 unprotected, and even little better than AW IMR's. I'm getting the full amp reading for Nanjg drivers confiured at 3.85A (reading 3.92A on fully charged PD's or Samsung's).
Just an fyi on all those observations and calculations above, SinkPAD's have real output advantages at higher amps - Match posted the test results and charts for aluminum vs. copper SinkPAD for XM-L2's, and it's quite clear running on copper SinkPAD's boosts output, but the boost ramps up the higher the amps go. What relic38 did, in upgrading a TR3T6, is do the test on a low driven setup where you won't see the biggest advantage of a SinkPAD - he's got 5.6A going to 3 LED's, about 1.9A each, and with U3 1C's, the difference will be neglible. In this setup, there is an advantage os SinkPAD's but only for running cooler, longer. For amps above 2.8A per LED, or better higher still, there are significant advantages in output, and it all makes sense because higher amps, heat sag stars kicking in, but the SinkPAD keeps it cooler, so by the time you hit 5 amps, the SinkPAD XM-L2 is 28% brighter than the XM-L2 on aluminum and still climbing.
Also, the difference between a T6 1A tint (or maybe 0D) and a U3 1C tint is probably not as high as you may think because of the tint differences.
Well, I got my first XM-L2 (T6 1C) from Fasttech (they threw it in my order for free; I guess they liked my Convoy C8 Review). I decided to put it in my TangsFire HD2010 from Tmart. Well what a nasty light to mod. The design of this thing just slaps you right in the face whenever it can. The reflector base is too wide; shorts on emitter wires. The driver holder doesn’t easily allow many driver options. Mine came with an O-ring on the pill; without it, nothing fits right. And it was broken so I had to replace it.
After lots of grinding, and twisting, and cursing, I managed to get an XM-L2 on copper in this light. Was it worth it? We’ll let the numbers decide…
Before: XM-L T6 1A, 16mm Al star, adhesive to pill. East-092 driver (DD), ~4.5A on Sanyo UR18650ZT (charged to 4.2V, 4.3V is well over 5.5A), no spring mods (driver spring slightly melted)
897 lm OTF at start, 795 lm OTF at 30s
Throw: 47kcd
After: XM-L2 T6 1C, trimmed 20mm copper SinkPad, Silver compound to pill. East-092 driver (DD), ~4.6A on TF Flame 26650 (Only 3.7A on Sanyo ZT at 4.3V), braided driver and tailcap springs
1276 lm OTF start, 1235 lm OTF at 30s
Throw: 81kcd
Wow! Impressive! This is a real performer, especially for a T6 bin. Seems to line up well with Tom E’s results too (his is driven less, higher bin, UCL lens, slightly more output and throw).
The higher Vf of the XM-L2 is obvious when when running on copper. When I tried running a 26650 on my XM-L HD2010 it goes over 6A; without braiding the springs! It would smoke in no time with braid added.
The beam is even nicer than before (when the emitter is centered). I can detect the slight improvement in tint (1A to 1C).
OK, time to order some more XM-L2’s! J)
It could be a super easy mod… if the light was designed a little better. |(
Once it is done though, it’s worth it. Trust me. This thing challenges my BTU Shocker. The Shocker wins, but when you consider the size difference, unless you absolutely need 120kcd+ of throw, the modded HD2010 goes in your pocket every time. Edit: Wow, that’s a lot of commas; I’m going to bed now.
Now, once I go to XM-L2 U2 on copper in the BTU, things will be different
Thank you relic38! Was beginning to think I was only one doing these type of mods and I was going nuts in seeing these numbers... Confirmation!! Yes, the HD 2010 is a challenge - been re-working mine but it's been getting low priority - want to swap the KD V2 for a programmable Nanjg and maybe go 4.5A with it, and did work out seating the reflector a bit better - dremel/grinding down on the backside flat surface to add clearance for the connectors. You know, was thinking instead of using 22 gauge wire, why not try two thinner wires to each contact, maybe 28's? Clearance will be lower and should be maybe same amount of wire thickness.
Also, on my original build, I put in a P60 pill press fit into the alum pill. Well, did a poor job of measuring depth, so new setup the driver sits too high, and using a KK ICR, there's a depression in the neg. end of the battery from the force over time. So, planning on cutting out the old P60 and replacing it with new one but seating it deeper. I got a stock HD 2010 to use for reference, so should be better this go-around.
Oh - the XM-L2 T6 0D's are looking good - more slightly bluish tint but saw excellent throw #'s in a C8 build, little better than my XM-L2 U2 1C build, lumens maybe slightly less, very close though - very similar builds at 3.85A, same reflector, SinkPAD of course. The XM-L2 U2 1A's though is probably the best around for pure brightness.
yesterday re reflowed an xml2 u2 to the sinkpads the led worked fine BUT ,even with direct drive with AW IMR and samsung 20R,i cant get more than 3-3,5 amps,same result with driver that they give 4,5amps to xml1 in this they give much much lower,so i guess i did something wrong with the reflow !?
i’ve tryed 2 sinkpads so its no the board and one strange thing ,i get diferent measures from the 4 pads of the sinkpad,my meter works fine.
any suggestion?
How did I miss this post! Awesome numbers. Get that de-domed XP-G2 R5 on a SinkPAD! Oh, also an XP-E2 would probably do even better, de-domed might be a pencil beam. I'm thinking if I de-dome the XP-G2 now on a SinkPAD in a T20, could probably get 150 kcd.
You didn’t do anything wrong. The reflow went fine.
The XM-L2 has a higher Vf than the XM-L1 so will always pull less current in direct-drive than a comparable XM-L1. Additionally, the SinkPAD keeps the junction much cooler, further raising the Vf of the emitter.